- The model is loaded from the from the Library
- The image is captured by the device camera
- The image is converted to a tensor
- The tensor is passed to the model
- The model outputs the predicted object
- The first prediction takes longer time to be completed because the model is loaded from the library for the first time
- Could not cache the model using service worker because it depends on other online resources
- Passess 100% of the lighthouse audit except for the Performance because the model is loaded from the library